Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Did HYDRA play perfect chess? Impossible!

Author: gerold daniels

Date: 15:24:12 07/01/05

Go up one level in this thread


On July 01, 2005 at 14:28:30, Terry Giles wrote:

>On July 01, 2005 at 14:05:11, Rolf Tueschen wrote:
>
>>On July 01, 2005 at 13:41:05, Otello Gnaramori wrote:
>>
>>>On July 01, 2005 at 09:18:09, Rolf Tueschen wrote:
>>>
>>>>Since my perfect chessmachine ROLF played only 75% of the HYDRA moves against
>>>>Adams, you can see how far away HYDRA is still from perfect play. That is why
>>>>it's still making sense if more GM play show events.
>>>
>>>It's a path near to perfect, perfection is a "chimera" , but it's stronger than
>>>strongest humans listening to Mr. Adams words...that is why it's so damn
>>>difficult to beat:
>>>
>>>From Adams interview at chessbase
>>>...
>>>"I think it proves that Hydra is a much stronger ‘player’ than any other
>>>computer in the world. We may not be able to measure its strength in Elo, but it
>>>is huge. I also suspect Hydra is stronger than any other human opponent. Okay,
>>>it has to be proved in the future, but this is my impression at the moment and I
>>>suspect it is accurate. I mean from my point of view I don’t think I played
>>>terribly. I did my best and it just wasn’t good enough."
>>>
>>>
>>>w.b.r.
>>>Otello
>>
>>
>>Hi Otello,
>>
>>what Adams says in the interview is under the medication of 10000 he got for his
>>draw. Adams doesn't know the computerchess world. Certainly he never heard of my
>>perfect chess machine.
>>
>>Debating the Elo of HYDRA doesn't make sense because it hasn't played much games
>>yet. On playchesscom HYDRA wasn't successful at all. Now you cannot begin to
>>calculate the Elo for - say - Sturday night or Monday. Elo is either for the
>>whole of it or nothing.
>>
>>If a human chessplayer had beaten Adams with 5,5 to 0,5 THEN this human
>>chessplayer had a tremendous Elo performance. But if you let Adams play a 6
>>games showevent against HYDRA, a completely unknown machine, and then calculate
>>Elo number with human touch so to speak. There is no such number.
>>
>>My tiny little program ROLF has Elo of over 30000, because it won all the games
>>it played. No wonder because it plays perfect chess with its 640 processors!
>>
>>That should give you a light impression of the power of my program. HYDRA and DB
>>and all the rest are millions of lightyears below - because as you know Elo
>>can't be calculated with basic retangular maths. So a three times stronger
>>player cant claim Elo 6000 if it has beaten a 2000 player. Know what I mean?!
>>
>>Elo 30 000 (In words thirty thousand) is as if you would leave the solar system
>>of this universe and enter new worlds, so to speak. I hope I could clarify a
>>bit.
>>
>>Rolf, Actually from Outer Space
>
>
>Hi Rolf,
>
>Working within the domain of unrestrained analytic pan-dimensional quantum
>hyper-maths and utilising Carol Vordermanian normalising vector indices I
>recently evaluated that the maximum theoretically attainable elo value to be
>somewhere in the region of 29700 +/- 200. I therefore regrettably, have to
>harbour some doubt as to the veracity of your claim.
>
>Terry (a bipedal carbon based life form)
>
>;-)

i think your tolerance of +/- 200 i a little on the high side.
gerold.



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.